(IJCSIS) International Journal of Computer Science and Information Security, Vol. 7, No.

2, 2010

Multi-objective Geometric Programming Problem with Weighted Mean Method
A. K. Ojha
School of Basic Sciences IIT Bhubaneswar Orissa, Pin-751013, India . Abstract-Geometric programming is an important class of
optimization problems that enable practitioners to model a large variety of real-world applications, mostly in the field of engineering design. In many real life optimization problem multi-objective programming plays a vital role in socio economical and industrial optimizing problems. In this paper we have discussed the basic concepts and principle of multiple objective optimization problems and developed geometric programming (GP) technique to solve this optimization problem using weighted method to obtain the non-inferior solutions.

K. K. Biswal
Department of Mathematics C.T.T.C, Bhubaneswar Orissa, India-751024 . functions. In this case, it is not suitable to use any singleobjective programming to find an optimal compromise solution. Biswal[4] has studied the optimal compromise solution of multi-objective programming problem by using fuzzy programming technique [20, 21]. In a recent paper Islam and Ray[8] find the Pareto optimal solution by considering a multi-objective entropy transportation problem with an additional restriction with generalized fuzzy number cost. In this paper we have developed the method to find the compromise optimal solution of certain multi-objective geometric programming problem by using weighting method. First of all the multiple objective functions transformed to a single objective by considering it as the linear combination of the multiple objectives along with suitable constants called weights. By changing the weights the most compromise optimal solution has been arrived by using GP techniques. The organization of the paper is as follows: Following the introduction, Formulation of multi-objective GP and corresponding weighting method have been discussed in Section 2 and 3. The duality theory has been discussed at Section 4 to find the optimal value of the objective function and the illustrative examples have been incorporated at Section 5 to understand the problem. Finally, at Section 6 some conclusions are drawn from the discussion. 2 FORMULATION OF MULTI-OBJECTIVE GEOMETRIC PROGRAMMING
A multi-objective geometric programming problem can be defined as: Find

Keywords- Geometric programming, optimization, weighted
method, duality theory, non-inferior solutions.

1 INTRODUCTION Geometric Programming(GP) problems have wide range of application in production planning, location, distribution, risk managements, chemical process designs and other engineer design situations. GP problem is an excellent method when decision variable interact in a non-linear, especially, in an exponential fashion. Most of these GP applications are posynomial type with zero or few degrees of difficulty. GP problem whose parameters, except for exponents, are all positive are called posynomial problems, where as GP problems with some negative parameters are refereed to as signomial problems. The degree of difficulty is defined as the number of terms minus the number of variables minus one, and is equal to the dimension of the dual problem. When the degree of difficulty is zero, the problem can be solved analytically. For such posynomial problems, GP techniques find global optimal solutions. If the degree of difficulty is positive, then the dual feasible region must be searched to maximize the dual objective, while if the degree of difficulty is negative, the dual constraints may be inconsistent. For detail discussions of various algorithms and computational aspects for both posynomial and signomial GP refers to Beightler et al.[2], Duffin et al.[6], Ecker[7] and Phillips et al.[13]. From early 1960 a lot of research work have been done on these GP problems [3, 5, 9, 11, 14, 16, 17, 18]. Mainly, we use GP technique to solve some optimal engineering design problems[1] where we minimize cost and/or weight, maximize volume and/or efficiency etc. Generally, an engineering design problem has multiple objective

x = ( x1 , x 2 ,...., xn ) T
Tk 0 n a

so as to

min : g k 0 ( x) = ∑ C k 0t ∏ x j k 0 tj , k = 1,2,..., p
t =1 j =1

(2.1)

subject to

g i ( x) = ∑ Cit ∏ x j itj ≤ 1, i = 1,2,..., m
d t =1 j =1

Ti

n

(2.2) (2.3)

x j > 0 , j = 1,2,..., n

where Ck0t for all k and t are positive real numbers and ditj and ak0tj are real numbers for all i,k, t, j.

82

http://sites.google.com/site/ijcsis/ ISSN 1947-5500

(IJCSIS) International Journal of Computer Science and Information Security, Vol. 7, No. 2, 2010

Tk0 = number of terms present in the kth objective function. Ti= number of terms present in the ith constraint. In the above multi-objective geometric program there are p number of minimization type objective function, m number of inequality type constraints and n number of strictly positive decision variables. 3 WEIGHTING METHOD OF MULTI-OBJECTIVE FUNCTIONS: The weighting method is the simplest multi-objective optimization which has been widely applied to find the noninferior optimal solution of multi-objective function within the convex objective space. If f1(x), f2(x),… ,fn(x) are n objective functions for any vector x = ( x1 , x 2 ,...., x n ) then we can define weighting
T

4.DUAL FORM OF GPP The model given by (3.2), (3.3) and (3.4) is a conventional geometric programming problem and it can be solved directly by using primal based algorithm for non linear primal problem or dual programming [12]. Methods due to Rajgopal and Bricker [15], Beightler and Phillips[1] and Duffin et al.[6] projected in their analysis that the dual problem has the desirable features of being linearly constrained and having an objective function with structural properties with suitable solution. According to Duffin et al.[6] the model given by (3.3) can be transformed to the corresponding dual geometric program as:

⎛w C max ∏ ⎜ k k 0t ⎜ w w t =1 ⎝ 0t
T0

⎞ ⎟ ⎟ ⎠

w0 t

⎛ wi 0 C ∏∏ ⎜ w it ⎜ i =1 t =1 ⎝ it
m Ti

⎞ ⎟ ⎟ ⎠

wit

∏ λ (w
t =1

Ti

it

) λ ( wit )
(4.1)

method for their optimal solution as defined below: Let W = ⎨w : w ∈ R , wi > 0,
n

⎧ ⎩

∑w
j =1

n

j

⎫ = 1⎬ to be the set of ⎭

subject to

∑w
t =1 m Ti

T0

0t

=1 wit + ∑∑ d itj wit = 0 , j=1,2,…,n
i =1 t =1 m Ti

non-negative weights. The weighted objective function for the multiple objective function defined above can be defined as P(w) where P ( w) = min
x∈X

∑∑ a
i =1 t =1

itj

∑w
j =1

n

j

f j ( x)

(3.1)

wit ≥ 0 ∀ t , j

It must be made clear, however, that if the objective space of the original problem is non-convex, then the weighting method may not be capable of generating the efficient solutions on the non-convex part of the efficient frontier. It must also be noted that the optimal solution of a weighting problem should not be used as the best compromise solution, if the weights do not reflect the decision maker's preferences or if the decision maker does not accept the assumption of a linear utility function. For more detail about the weighted method refer[10]. Based on the importance of the p number of objective functions defined in (2.1) the weights w1,w2,…,wp are assigned to define a new min type objective function Z(w) which can be defined as

∑w
k =1

p

k

= 1, wk > 0, k = 1,2,..., p

Since it is a usual dual problem then it can be solved using method relating to the dual theory. 5 NUMERICAL EXAMPLES For illustration we consider the following examples. Example: 1 Find x1, x2, x3, x4 so as to min : g10 ( x) = 4 x1 + 10 x2 + 4 x3 + 2 x4 (5.1)

max : g 20 ( x) = x1 x2 x3
subject to

(5.2) (5.3) (5.4)

min : Z ( x) = ∑ wk g k 0 ( x)
k =1

p

=

n ⎛ Tk 0 a wk ⎜ ∑ C k 0t ∏ x j k 0 tj ∑ ⎜ t =1 k =1 j =1 ⎝ p

⎞ ⎟ ⎟ ⎠

(3.2)

=
Ti

∑∑ w C ∏ x
k =1 t =1 n d k k 0t j =1 j =1

p Tk 0

n

x x + ≤1 x x 100 ≤1 x1 x2 x3 x1 , x2 , x3 , x4 > 0

2 1 2 4

2 2 2 4

ak 0 tj j

Now the problem can be rewritten as

min : g10 ( x) = 4 x1 + 10 x2 + 4 x3 + 2 x4
(3.3)

(5.5) (5.6) (5.7) (5.8) (5.9)

subject to

∑ Cit ∏ x j itj ≤ 1 , i = 1,2,..., m
t =1

min : g ( x) = x x x
/ 20

−1 −1 −1 1 2 3
−2 4 −1 3

Subject to x x

x j > 0, j = 1,2,..., n
where

∑w
k =1

p

k

= 1, wk > 0, k = 1,2,..., p

(3.4)

+ x x ≤1 100 x x x ≤ 1 x1 , x2 , x3 , x4 > 0
2 2 −1 −1 1 2

2 −2 1 4

83

http://sites.google.com/site/ijcsis/ ISSN 1947-5500

(IJCSIS) International Journal of Computer Science and Information Security, Vol. 7, No. 2, 2010 w05 w11 w12 w21 Z Introducing weights for the above objective functions a new 0.0010217 0.051051 0.014213 0.3319702 8.80776 objective function is formulated as
− − Z ( x) = w1 (4 x1 + 10 x2 + 4 x3 + 2 x4 ) + w2 ( x1−1 x2 1 x3 1 )

0.0004543 0.000265092 0.000170424 0.000113614

0.051080 0.05109 0.051095 0.051098

0.014221 0.0142237 0. 014225 0.0142259

0.3319701 0.3329799 0.3331061 0.3331818

17.60555 26.40333 35.201110. 43.99888

(5.10) subject to x x

+ x x ≤1 100 x x x ≤ 1 x1 , x2 , x3 , x4 > 0 where w1 + w2 = 1, w1 , w2 > 0
2 2 −1 −1 1 2

2 −2 1 4

−2 4 −1 3

(5.11) (5.12)

(5.13) This problem is having degree of difficulty 3. The problem is solved via the dual programming [6]. The corresponding dual program is:

Using primal dual relationship the corresponding primal solution are given in the following Table. Table-2 Primal solution:
w1 0.1 0.2 0.3 0.4 w1 0.9 0.8 0.7 0.6 0.5 x1 5.084055 5.084055 5.084055 5.084055 5.084055 x2 2.682555 2.682555 2.682555 2.682555 2.682555 x3 7.332315 7.332315 7.332315 7.332315 7.332315 x4 5.748367 5.748367 5.748367 5.748367 5.748367 Z 8.80776 17.60555 26.40333 35.20111 43.99888

⎛ 4w ⎞ max : V ( w) = ⎜ 1 ⎟ ⎜w ⎟ w ⎝ 01 ⎠ ⎛ w2 ⎞ ⎜ ⎜w ⎟ ⎟ ⎝ 05 ⎠
w05

w01

⎛ 10w1 ⎞ ⎜ ⎜ w ⎟ ⎟ ⎝ 02 ⎠
w12

w02

⎛ 4w1 ⎞ ⎜ ⎜w ⎟ ⎟ ⎝ 03 ⎠

w03

⎛ 2w1 ⎞ ⎜ ⎜w ⎟ ⎟ ⎝ 04 ⎠
)

w04

⎛ 1 ⎞ ⎜ ⎜w ⎟ ⎟ ⎝ 11 ⎠

w11

⎛ 1 ⎞ ⎜ ⎜w ⎟ ⎟ ⎝ 12 ⎠

(w11 + w12 )(w

11 + w12

100 w21
(5.14)

0.5

subject to w01 + w02 + w03 + w04 + w05 = 1

where the minimum values of g10 = 87.98776 and g20 = 0.01 Example: 2 Find x1, x2, x3, x4 in order to
− − min : f1 ( x) = x1−1 x2 1 x3 1

w01 − w05 + 2w11 − w21 = 0 w02 − w05 − 2 w12 − w21 = 0 w03 − w05 − w21 = 0 w04 − 2w11 − 2w12 = 0
w1 + w2 = 1

(5.15) (5.16) (5.17) (5.18) (5.19)

− − − min : f 2 ( x) = x1−1 x2 3 x3 5 + x1−1 x2 1

subject to x x x + x 2 x3 ≤ 6
2 1 2 3

w01 , w02 , w03 , w04 , w05 , w11 , w12 , w21 ≥ 0
w1 , w2 > 0
By considering different values of w1 and w2 the dual variables, corresponding maximum value of dual objective are given in the following Table. Table-1 Dual solution:
w1 0.1 0.2 0.3 0.4 0.5 w2 0.9 0.8 0.7 0.6 0.5 w01 0.2308894 0.2310206 0.2310643 0.2310862 0.2310993 w02 0.3045667 0.3044397 0.3047974 0.3048263 0.3048436 w03 0.3329927 0.3331819 0.3332450 0.3332765 0.3332955 w04 0.1305293 0.1306035 0.1306282 0.1306406 0.1306480

x1 x3 ≤ 1 x1 , x2 , x3 > 0

Using the weights the above objective function can be reduced to the new objective function as:
− − − − − Z ( x) = w1 ( x1−1 x2 1 x3 1 ) + w2 ( x1−1 x2 3 x3 5 + x1−1 x 2 1 )

(5.20)

(5.24) In this problem the degree of difficulty is 2 and it can be solved by using duality theory as given by

x x x + x 2 x3 ≤ 6 x1 x3 ≤ 1 x1 , x2 , x3 > 0 where w1 + w2 = 1 w1 , w2 > 0
2 1 2 3

(5.21) (5.22) (5.23)

⎛w ⎞ max : V ( w) = ⎜ 1 ⎟ ⎜w ⎟ w ⎝ 01 ⎠
w11

w01

⎛ w2 ⎞ ⎜ ⎜w ⎟ ⎟ ⎝ 02 ⎠
w12

w02

⎛ w2 ⎞ ⎜ ⎜w ⎟ ⎟ ⎝ 03 ⎠

w03

⎛ 1 ⎞ ⎛ 1 ⎞ ( w11 + w12 ) ⎜ ⎟ ⎟ ⎜ ⎜ 6 w ⎟ ⎜ 6 w ⎟ (w11 + w12 ) ⎝ 11 ⎠ ⎝ 12 ⎠ subject to w01 + w02 + w03 = 1

(5.25)

84

http://sites.google.com/site/ijcsis/ ISSN 1947-5500

(IJCSIS) International Journal of Computer Science and Information Security, Vol. 7, No. 2, 2010

− w01 − w02 − w03 + w11 + w21 = 0 − w01 − 3w02 − w03 + w11 + w12 = 0 − 2w01 − 5w02 + 2w11 + w12 + w21 = 0 w01 , w02 , w03 , w11 , w12 , w21 ≥ 0
w1 , w2 > 0
For different values of w1, w2 the dual variables and the corresponding maximum values of dual objective is obtained as given in the Table Table-3 Dual solution: w01 w02 0.2085711 0.5276192 0.3640122 0.4952513 0.604162 0.6959958 w12 1.055235 0.9239918 0.8364992 0.7638920 0.7026694 0. 4239919 0.3364992 0.2638920 0. 2026694 w21 0 0.076008 0.1635008 0.2361080 0.2973305

6 CONCLUSIONS By using weighted method we can solve a multi-objective GPP as a vector minimum problem. A vector-maximum problem can be transformed as a vector minimization problem. If any of the objective function and/or constraint does not satisfy the property of a posynomial after the transformation, then we use any of the general purpose nonlinear programming algorithm to solve the problem. We can also use this technique to solve a multi-objective signomial geometric programming problem. However, if a GPP has either a higher degree of difficulty or a negative degree of difficulty, then we can use any of the general purpose nonlinear programming algorithm instead of a GP algorithm. REFERENCES

w1 0.1 0.2 0.3 0.4 0.5

w2 0.9 0.8 0.7 0.6 0.5 w11 1.00

w03 0.2638096 0.2119959 0. 1682496 0. 1319460

[1] [2] [3]

[4]

0. 1013347
[5]

Z 0.1642316
[6]

0.9239919 0.8364992 0.7638920 0. 7026695

0.1831441 0.2019177 0.2206914 0.2394650
[9] [7] [8]

The corresponding primal solution is given in the following Table: Table-4 Primal Solution x1 x2 x3 Z w1 w1 0.1 0.9 2.527860 8.217575 0.3748833 0.1642316 0.2 0.3 0.4 0.5 0.8 0.7 0.6 0.5 2.620746 2.620745 2.620747 2.620747 7.862237 7.862236 7.862240 7.862242 0.3815708 0.3815709 0.3815707 0.3815705 0.1831441 0.2019177 0.2206914 0.2394650

[10] [11] [12]

[13] [14]

[15]

Ideal solution of the two objective functions are given below: f1 = 0.33333; x1 = 236.9322; x2 = 710.7964; x3 = 0.0042206 and f2 = 0.1421595; x1 = 2.148558; x2 = 9.82199; x3=0.3490711.

[16]

[17]

[18]

C.S.Beightler and D.T.Phillips: Applied Geometric programming, John wiley and Sons, New York., 1976. C.S.Beightler and D.T.Phillip, D.J.Wilde: Foundations of optimization, Prentice-Hall, New Jersy, 1979. H.P.Bensonand, G.M. Boger: Multiplicative programming problems analysis and efficient point search heuristic, Journal of optimization theory and applications 94, 487-510, 1997. M.P. Biswal: Fuzzy Programming technique to solve multi-objective geometric programming problem, Fuzzy sets and systems 51,67-71, 1992. C.Chu and D.F.Wong: VLSI Circuit performance optimization by geometric programming, Annals of Operations Resarch 105, 37-60, 2001. R.J.Duffin, E.L.Peterson and C.M.Zener: Geometric Programming Theory and Application, Wiely, New York , 1967. J.G.Ecker: Geometric programming Methods, computations and applications. SIAM Review 22, 338-362, 1980. S.Islam and T.K.Ray: A new fuzzy multi-objective programming: Entropy based geometric programming and its applications of transportation problems, Europian Journal of Operational Research, 173, 387-404, 2006. C. Kao and S.T.Liu: Predicting bank performance with financial forecasts:A case of Taiwan commercial banks, Journal of Banking and Finance 28, 2353-2368 , 2004. G.P.Liu, J.B.Yang and Whidborne: Multi-objective optimization and contral, PHI, EEE, 2006. S.T.Liu: Geometric Programming with parametric uncertainty, Europian Journal of Operation Research 168, 345-353, 2006. E.L.Peterson: The fundamental relations between geometric programming duality, Parametric programming duality, and ordinary Lagrangian duality, Annals of Operations Research 105,109-153, 2001. D.T.Phllips, C.S.Beightler: A technical state of the art survey.AIIE Transactions 5, 97-112, 1973. J.Rajgopal D.L Bricker: Posynomial geometric programming as a special case of semi-infinite linear programming, Journal of Optimization Theory and Applications 66, 455-475, 1990. J.Rajgopal and D.L.Bricker: Solving Posynomial geometric programming problems via generalised linear programming, Computational Optimization and Applications 21, 95-109,2002. C.H.Scott and T.R.Jefferson: Allocation of resources in project management, International Journal of Systems Science 26, 413-420, 1995. H.D.Sherali: Global Optimization of nonconvex polynomial programming problems having rational exponents, Journal of global Optimization 12, 267-283, 1998. S.B.Sinha, A.Biswas and M.P.Biswal: Geometric Programming problem with negative degrees of difficulty, Europian Journal of Operations Research 28, 101-103, 1987.

85

http://sites.google.com/site/ijcsis/ ISSN 1947-5500

(IJCSIS) International Journal of Computer Science and Information Security, Vol. 7, No. 2, 2010
[19] B.M.Worrall, M.A.Hall: The analysis of an inventry central model using posynomial geometric programming, International Journal of Production Research 20, 657-667, 1982. [20] H.J.Zimmermann : Fuzzy programming and linear programming with several objective functions, Fuzzy sets and systems 1, 46-55, 1978. [21] H.J.Zimmermann: Fuzzy set theory and its applications, 2nd ed. Kluwer Academic publishers, Dordrecht-Boston, 1990. AUTHORS PROFILE

K.K.Biswal: Mr.K.K.Biswal received M.Sc.(Mathematics) from Utkal University in 1996. Currently he is a lecturer in Mathematics at CTTC, Bhubaneswar, India. He is performing research works in Geometric Programming. He is served more than 7 years in different colleges in the state of Orissa. He has published 2 research papers in different journals.

Dr.A.K.Ojha: Dr.A.K.Ojha received a Ph.D(mathematics) from Utkal University in 1997. Currently he is an Asst.Prof. in Mathematics at I.I.T. Bhubaneswar, India. He is performing research in Nural Network, Geometric Programming, Genetical Algorithem, and Particle Swarm Optimization. He has served more than 27 years in different Govt. colleges in the state of Orissa. He has published 22 research papers in different journals and 7 books for degree students such as: Fortran 77 Programming, A text book of modern algebra, Fundamentals of Numerical Analysis etc.

86

http://sites.google.com/site/ijcsis/ ISSN 1947-5500